API: consistent NaN treatment for pyarrow dtypes #24297

Job	Run time
Linux-32-bit	18m 32s
Linux-Musl	15m 22s
Pyodide build	23m 50s
python-dev	0s
windows-latest actions-312.yaml	36m 56s
macos-14 actions-310.yaml	20m 35s
macos-13 actions-312.yaml	31m 32s
macos-14 actions-313.yaml	21m 17s
macos-14 actions-311.yaml	21m 32s
macos-13 actions-310.yaml	34m 38s
macos-13 actions-313.yaml	43m 30s
ubuntu-24.04 actions-310.yaml	48m 8s
windows-latest actions-313.yaml	43m 24s
windows-latest actions-310.yaml	36m 31s
windows-latest actions-311.yaml	37m 21s
macos-13 actions-311.yaml	32m 53s
ubuntu-24.04-arm actions-311.yaml	35m 20s
macos-14 actions-312.yaml	19m 14s
ubuntu-24.04 actions-311.yaml	42m 35s
Freethreading	21m 10s
Minimum Versions	23m 17s
ubuntu-24.04-arm actions-313.yaml	35m 39s
Numpy Dev	16m 44s
Downstream Compat	6m 3s
ubuntu-24.04-arm actions-310.yaml	36m 53s
Locale: zh_CN	23m 50s
Locale: it_IT	24m 1s
Future infer strings	46m 33s
Pyarrow Nightly	21m 16s
ubuntu-24.04 actions-312.yaml	45m 43s
ubuntu-24.04 actions-313.yaml	47m 17s
Pypy	1m 51s
Future infer strings (without pyarrow)	38m 29s
ubuntu-24.04-arm actions-312.yaml	36m 52s
	16h 28m 48s

Provide feedback